Role of Thyroid Ultrasound in the Diagnostic Evaluation of Thyroid Nodules: A Systematic Review

Abstract

Thyroid nodules are highly prevalent, with the majority being benign and only a small percentage turning out to be malignant. Accurate and non-invasive differentiation between benign and malignant nodules is essential for appropriate clinical management. This systematic review evaluates the diagnostic performance of thyroid ultrasound in comparison with standardized risk stratification systems, advanced imaging techniques such as elastography, and emerging artificial intelligence-assisted methods. A comprehensive literature search was conducted across PubMed, Scopus, Web of Science, and Google Scholar from 2010 to 2024, focusing on studies that reported sensitivity, specificity, and predictive values using fine-needle aspiration cytology (FNAC) or histopathology as the reference standards. Data extraction and quality assessment were performed by two independent reviewers following PRISMA guidelines. Thirty-two studies were included, revealing that conventional ultrasound has an average sensitivity and specificity of approximately 88% and 86%, respectively. Diagnostic performance was notably enhanced with the application of risk stratification systems like TIRADS, elastography, and AI-based tools. However, diagnostic accuracy varied depending on operator expertise, patient population, and ultrasound methodology. Ultrasound alone was found to be insufficient in confidently excluding malignancy in indeterminate cases, justifying the continued use of FNAC in selected patients. Overall, thyroid ultrasound, particularly when combined with classification systems and AI, proves to be a cost-effective and valuable primary screening tool. Further multicenter studies are recommended to promote standardization, validate findings, and enhance integration of AI in clinical practice.